Transcript: They Know What You Read

Listen to Podcast Download Transcript PDF

Recorded at the Digital Book World Conference + Expo 2015


  • Micah Bowers, Co-Founder and CEO, Bluefire.
  • David Burleigh, Director of Marketing & Communication, OverDrive.
  • Jared Friedman, Co-founder & CTO, Scribd.
  • Michael Tamblyn, President & Chief Content Officer, Kobo

For podcast release Monday, January 19, 2015

KENNEALLY: The owners of e-book platforms now have unprecedented and previously unattainable knowledge about how people – you – read. They literally see every time an e-book is opened, on what device it is opened, how fast it is read and whether passages or entire works are reread. Perhaps most dismaying of all for authors and publishers alike, they even know how many books that are bought are never cracked, let alone finished.

The great dramatist Jean Racine once observed, they are no secrets that time does not reveal. To the inescapable force of time, we may now add technology. And bear in mind that Racine was a master of tragedy, so we should wonder, if only for a moment, just how welcome we will all find this transparency.

This afternoon, representatives of global e-book platforms will offer revealing insights into what they know about what you read. We’ll hear about consumer reading behavior across multiple devices and whether fans of fiction read differently than nonfiction aficionados. We’ll also learn whether and to what extent reading of English goes on in non-English-speaking countries.

And to do all of that, I want to welcome my panelists. And I’ll move from my left out, and we’ll start with Michael Tamblyn, President and Chief Content Officer at Toronto-based Kobo. Michael, welcome.

TAMBLYN: Hi, Chris. Thanks for having us.

KENNEALLY: We’re delighted that you could join us. Founded in 2009 and acquired in 2011 by Tokyo-based e-commerce company Rakuten, Kobo has become a global leader in e-reading. Kobo delivers digital reading experiences to millions of users in 190 countries, offering one of the world’s largest catalogs, with millions of titles, world-class e-reading devices and top-ranking apps.

Prior to joining Kobo, Michael Tamblyn was the founding CEO of supply chain agency BookNet Canada, where he launched the national sales reporting service BNC SalesData. He also co-founded Canada’s first online bookstore, which was purchased by Indigo Books & Music in 1998.

And then, to Michael’s left, we have Jared Friedman. Jared, welcome.

FRIEDMAN: Thanks, Chris. It’s good to be here.

KENNEALLY: Jared is a co-founder and currently CTO of Scribd. And he coded the original version of Scribd during his junior year at Harvard, converting his dorm room into a makeshift datacenter. You weren’t alone in doing that, I bet. I’m sure there were a few datacenters there in Harvard Square.

Today Scribd – and Jared – is based in foggy San Francisco, where he’s building the world’s premier e-book subscription service. Since its 2013 launch, Scribd readers have clocked more than 17 million hours of reading time across all genres. Then, to Jared’s left, we have David Burleigh. David, welcome.

BURLEIGH: Thanks, Chris. Good to be here.

KENNEALLY: David Burleigh is Director of Marketing and Communications at OverDrive, the leading e-book and digital media platform for libraries, schools and retail. In this role, David leads the company’s brand marketing efforts, PR and social media in both B2B and B2C. He has 20 years of experience – in marketing global products and services in a variety of industries.

And then finally, at the very far end, Micah Bowers. Micah, welcome. Micah founded Bluefire Productions in 2001 and serves as its User Experience Director. Micah leads the Bluefire project teams and collaborates with clients on content strategy and experience design, and Bluefire Productions provides e-reading solutions to publishers, retailers and companies who create or distribute high-value content in support of their core businesses. And it’s important to describe what each of these particular men do in their working lives because we really are going to get a very broad, very comprehensive view of reading behavior across a variety of different channels and audiences.

And so I think one question that everybody has, because I bet you’re like me, you’ve got at least two devices with you in this room, so I want I start with a question that I believe, Jared, you have some information on, which is about how many people are reading on one device versus on multiple devices. Can you tell us about that?

FRIEDMAN: Sure. So Scribd is a multi-device ecosystem, and so we really try to encourage people to read on multiple devices. And when we looked at the results to see how many people actually were, we were really excited to see that actually more than half of readers read on at least two devices, and actually 10% of readers read on three or more devices in any given month, which is very exciting for us. And what’s also exciting is that the percentage of multi-device users is growing, so I think, as people become more familiar with e-books, with e-book reading services, they’re allowing it to sort of permeate their digital lifestyle. And as people get more devices into their homes, they’re reading on more and more devices.

KENNEALLY: Well, Michael, if I could ask you about that question, and a different twist, if you will, because you’ve got an e-reader, but you’re also available on apps. Do we know about behavioral differences on the reader versus the app?

TAMBLYN: Absolutely, and it’s fascinating, because we can either look at it in terms of our base of users who are e-reader users who have a device on the side. We have people who are app-only all the time. And then there’s some really interesting geographical variations as well.

But if you look at the whole base and you roll it up all together, because we have this great base of people using our devices, we have a much lower level of kind of multi-device use than Jared does. So we see more like, you know, kind of 15% to 20% are multi-device users. Most people are picking one and sticking with it. But when we looked at that, we then had to rerun all of the data for Japan, for example, because in Japan there are no e-readers, and it’s effectively an only smart phone and table market, and their multi-device use is much higher.

KENNEALLY: OK. David Burleigh, I think it’s important to distinguish what OverDrive does. It’s not only the market you serve, the libraries, not a consumer market, per se, a different type of consumer, but the way that OverDrive provides or delivers the book experience to that reader. Tell us about that.

BURLEIGH: Right. Well, it certainly is the consumer market. It’s just we deliver it through the library and school channels. So these are users that are acquiring – borrowing a book really – for a certain lending period that they’ll read on different devices. We provide apps for all the different devices. We supply Kindle as well, library e-books on their Kindle in the U.S. only, and then certainly on the desktop. So we’re seeing mobile use has gone up significantly, smart phones and tablets, especially. We can’t break out the e-reader specifically because the e-reader – the numbers for Kindle come from either the desktop or the tablet or the phone to get it on to Kindle.

KENNEALLY: Right. So what you’re seeing is what these other gentlemen also see as well, which is a proliferation of devices and presumably readers who are reading on all those devices or many of them.

BURLEIGH: The numbers that we have are closer to what Scribd is seeing, where we see about 40% of the users that are borrowing e-books are reading them on more than one device.

KENNEALLY: OK. Micah Bowers, again it’s a very distinctive offering that Bluefire has. Would you describe it better than I may have done for our listeners here, or for our audience here, and then address that question about multiple devices and uses across those devices?

BOWERS: Sure. Well, in the case of Bluefire Reader, it’s interesting because it’s not associated with any particular retail provider or any particular library. We have users that use the application that get content from many different varieties of sources, lots of different retailers, lots of different library systems, content that the user owns themselves, personal documents, etc. So we, it’s a very different – So while we may share similar types of users with the other folks at this table, it’s, I think, a much broader spread.

Now, in our particular case, we operate within a fairly open ecosystem in the sense that people who are using our Reader applications are very often using other applications to read those same books as well. So the fact that someone might be reading a particular book with Bluefire Reader on an iPad, they may be using a completely different application to use it on the desktop or they may be using an e-device from another provider, and they – well, we don’t provide e devices, but from a variety of – that might be associated with a retailer or might be a device they bought that’s not specific to a retailer. And so, while we have a good analytic picture based on users using our apps, we recognize that we only, perhaps, are seeing a snapshot of their total reading behaviors.

KENNEALLY: Well, and I think the audience should bear in mind that that’s true of everything that we’re going to hear this afternoon. We really can’t say that this is a complete picture of any particular reader or even of groups of readers.

The kinds of data that you’re all collecting is, of course, aggregated data. We want to reassure people that we’re not looking to see what you’re reading at night versus, well, actually we know what you’re reading at night, what types of genres, but not the particular titles. And we know that you’re hopefully not reading in the car, you’re listening to audiobooks in the car.

For example, Jared Friedman, Scribd service has an audio component as well. And you know therefore that different types of commuters are consuming different types of product as a result.

FRIEDMAN: Yeah. It’s very interesting to look at the genre distribution at different times of the day. People really do read very different books in the morning, in the afternoon and at night.

KENNEALLY: Well, indeed. And Michael Tamblyn, I know that Kobo released a reading habit, reading behavior report at the end of this year. And tell us what your findings were for the month of January. I think there were certain types of books. It’s not just when during the day, but when during the year that people are reading different types of things.

TAMBLYN: Well, again it’s a highly geographically sensitive area for us. So we do see swings between when things like romance become prevalent. There is a summer reading effect, where people go deeper into the genres. There is a kind of a New Year resolution effect, where people dig into nonfiction a little bit more.

But I think, for the most part, for the vast majority of users, their reading habits are relatively stable through the year. So taking a lot of the edge conditions out, a romance reader is a romance reader all the time. A science fiction and fantasy reader is dug into that genre fairly hard. And then what we’re interested in are kind of where do they overlap and where do those overlaps start to change?

KENNEALLY: And their behavior is different from reading affinity to different reading affinities. Again, Jared Friedman, I’ve seen some numbers to that point about New Year’s resolutions and so forth, a lot of people may have gotten some health and fitness books in their e-readers or they may be looking for self-help and fitness in their subscription services. But what’s their likelihood of actually finishing that?

FRIEDMAN: Yeah, so it’s perhaps an unfortunate truth about human nature, which we are able to observe in our data, which is that we do see a big spike in sort of aspirational reading, like what Michael was talking about in January, following the holidays. It’s particularly true in the health and fitness subgenre.

Unfortunately, as you might expect, we see that the completion rates of those books are quite low. So for something like a romance book, somebody who opens a romance book has something like a 30% chance of finishing it. Somebody who opens a health and fitness book has only about a 5% chance of actually finishing that book, so,

KENNEALLY: Well, those of us who are denizens of gyms and health centers know that the places are crowded in January, and they tend to open up a little bit more room in February and March, so we know what that’s like. And I was struck, there’s a particular, if I can call it, subgenre, one that appears to be popular if you see the proliferation of these kinds of centers but yet, for the reading world, it’s not so popular, or at least not finishing it, and that would be yoga. People just, that’s the least finished category of books.

M: Yes, it is.

M: But it’s like, and if I can build on that for one second, though, I think –


M: – when we look at completion, you have to look at fiction and nonfiction very differently. When people read fiction, they are reading an arc of a story. Generally, if they don’t get to the end, it’s because they don’t like it or they didn’t find value in it. Nonfiction tends to be dipped in and out of a lot more. People will pull the gist out of it. They’ll use it for reference. They’ll pick a particular sort of chapter or set of chapters, and that becomes where they extract value from it.

And if you ask them, was this a good book or did this book serve your needs, regardless of whether they get to the end, they’re quite likely to say, no, that was a fantastic book on yoga. I got the six things that I felt I needed to know about downward dog, Ananda, and (inaudible), whereas, for a work of fiction that doesn’t get finished, if they didn’t get to the end, it’s usually because they didn’t like it.

KENNEALLY: Well, right. And nonfiction is used and consumed in a different ways over different periods of time. I might need that chapter today and won’t be finishing my home project for six months from now, and I’ll go back to it at that point. David Burleigh, library readers have a very limited amount of time in contrast with Jared’s, if you will, all-you-can-eat buffet offering.


KENNEALLY: Does that make a difference at all?

BURLEIGH: Oh, it does, absolutely. So when you’re providing your books to OverDrive and into the public library space, the public library and the school or anybody else who’s lending that book, or e-book, in this case, has a set lending period, seven, 14 or 21 days, and then it locks up at the DRM.

So what we’ve found, and we can do this based on our browser-based reader, OverDrive Read, and we can tell how far you get in the book and how far, if you completed or not and page turns and things like that, how much time you spent, and that’s about 11% of our checkouts, whereas ePub is downloaded into our app, and we don’t track the reading behavior in there.

But we did look at the three most popular books of the year, Gone Girl, Unbroken and Fault in Our Stars and the different genres, and they were all about the same in terms of 30% made it halfway through these books. Only 6% to 11% finished these books. And that could be due to the length of the book – whether they liked it, but also probably more likely due to the lending period. So they just got through it, and then they just didn’t quite finish it. And then they would have to check it out again, or, in many cases, they can go buy it. And we see that behavior as well.

KENNEALLY: Right. And Jared, am I right that, being a subscription service, this notion of how quickly or how completely one reads a book isn’t so important to you?

FRIEDMAN: It’s not in a direct way based on (inaudible) business model, people have an unlimited amount of time to read books. And one interesting thing that we found is that people are actually surprisingly likely to return to a book even after a long intermission. So we find – we thought initially that, if people had abandoned a book for one month or more, that they’d hardly ever return to it.

KENNEALLY: It sounds like it’s orphaned, if you will.

FRIEDMAN: Right, exactly. But in fact 10% of people who have started a book, left it for a month or more, will actually return to it and eventually finish it.

KENNEALLY: And do you know if that’s true across genres and fiction and nonfiction?

FRIEDMAN: Well, to Michael’s point, it’s really a fiction measurement, because not all nonfiction is designed to be read start to finish.

KENNEALLY: OK. Micah Bowers, you don’t know about reading habits until somebody actually opens a book so, to a point that we’ll discuss later on about how many books get cracked, do you have any sense that sort of follows on what Jared and David were just saying about this completion rate and any conclusions we can draw from it?

BOWERS: Well, I would reiterate that we see the same, or I would reinforce that the difference between fiction and nonfiction is dramatic, and it really depends on the type of nonfiction. We as well see somewhere around a 30% completion rate of people who actually begin to read the book although, in our particular case, as we know that there are a certain number of users who are reading outside of our reading systems, we don’t know if they might be reading the first half of the book in Bluefire Reader and then finishing the other half somewhere else, so our total picture of completion rates are not complete, but we do see a surprising amount of completion of people. And it’s interesting, because you have this kind of difference. There’s the people who, if they’ve gotten past 10% in a book, their chance of finishing the entire book is much, much higher than the person who cracked the book. Right? So, and the person who gets halfway through is much, much, much higher to finish the entire book, which is logical, of course, but you do in fact see that in the statistics.

One of the interesting things to me, and a little off topic, but, is just the amount of – what has been really surprising to us, is just the amount of time people spend reading. I really thought that mobile was a thing where it was convenient, it’s on the go, you’re on the bus, you can read here and there, but we really found that people read a really surprising amount of time, whereas something like – according to our stats, people that – of any given users on a given day, we have about 35% read for over an hour. And then we have about 17% that read over two hours in a day, so that’s a pretty significant chunk of people that are reading for significant periods of time.

KENNEALLY: And that speaks to a question which everybody here cares about, which is commitment and engagement. It’s not just about measuring these things for measurement’s sake but to try to draw some inferences about what the books mean to people and how we can give more meaning to them in their reading experience. Michael Tamblyn, what do you think about that?

TAMBLYN: Well, that, for us, is a lot of the reason why this information is interesting. We’re mostly concerned about how do we use it to be better booksellers? You know, people who bought this also bought that is interesting, but getting a sense of whether you were engaged with a book, whether you like an author, is this a book that you stay up through the night in order to finish, you know, these are things that help us, from a recommendation perspective, figure out whether we should put that next book in front of you from that same author, how quickly should we feed you the next book in a series, or should we walk you towards another author, where the sort of reading and engagement and completion profiles look similar to the one that you’re just leaving? And all of that means someone more likely to be happy with the next book that we give them.

KENNEALLY: Right. And Kobo very much an international player in all of this. Give us a sense of the geographic differences, the national differences, where they are and maybe what surprises you about them.

TAMBLYN: Oh, it’s fascinating. You see both the genres that people are most engaged with change from country to country. U.S., Canada, U.K. are romance territories first. France is much – you see higher completion rates for mystery, and the same for Italy. And you see differences in the percentage likelihood that someone’s going to finish a book. French and Italian and Dutch readers are just more committed. You see numbers like, averages like 60%, 70%, 71% in EU countries, whereas the U.S. is the most fickle audience in the world. We see romance completion rates of like 44% here.

So people are, in some countries, just much more likely to dig in and stay with a book, and not in others. And then you, you know, a completely different world in Japan, where the books are often short. The reading behavior around manga and comic books looks completely different than fiction here. And people burn through them at the rate of multiple books a day.

KENNEALLY: Right. And while that confirms some suspicions, perhaps even some stereotypes, about national habits, there was one surprise, which I’d like you to comment on. And that is literary fiction. We think of the French, and they’re very proud of their literary tradition, which they safeguard very strongly. How literary are they?

TAMBLYN: The French reading public love trashy books just as much as anyone else.

KENNEALLY: That’s good to hear.

TAMBLYN: No, so we can hold our heads high, knowing that we are no more and no less deep in love with genre than any of the other sort of members of those great European reading superpowers.

KENNEALLY: Well, and Jared Friedman, what about that? Do we see some geographic differences that are important to note here?

FRIEDMAN: We do, for sure. So Scribd is strongest in English language content. So what we’ve really been amazed at is the amount of English language reading activity all over the world, particularly in geographies that the book publishing industry perhaps has underserved in the past. So we see a tremendous amount of reading activity in Southeast Asia, Singapore, Malaysia, Indonesia. And we see huge, huge genre differences there. So in the U.S., fiction, in particular genre fiction, is quite big. But in Southeast Asia, we see a lot more nonfiction, self-help, education, people really reading to improve themselves, to achieve a goal. In Western Europe, we see a lot more cooking and travel. An Italian is three times as likely to read a cookbook as an American. And when it comes to romance and erotica, actually the highest, highest index for us is Southeast Asia for erotica. I don’t know if you’re seeing the same thing, Michael. (laughter)

TAMBLYN: The stories I can tell.

KENNEALLY: Well, he may not want to tell you, exactly. You know, Michael brought up recommendations, which I’m sure is an important issue for you, Jared. And there are interesting associations there. For example, as I understand it, romances, people who read romances are also looking for which kind of books?

FRIEDMAN: So we see actually some really interesting overlaps between romances and sort of home and garden genres, like sewing, knitting, cooking, that actually appeal to a similar demographic, even though topically they’re entirely unrelated.

KENNEALLY: Right. And David Burleigh, at OverDrive, where again your concern is in libraries, that’s a great source traditionally for discovery. Point to some findings in your data that reveal something about that.

BURLEIGH: Sure. We know the library has always been, for hundreds of years, a great place to discover new books. And it’s no different for the digital world, where we’ve had some data, some surveys asking people if they go looking for a book, an e-book, specifically, do they just leave if they can’t find it or if it’s unavailable because it’s been checked out, or get something else while they’re there? We found that 76% are actually looking for something, will find something else, and then they’ll look for something else while they’re there.

We also know that, even if somebody has borrowed a book, 35% of them will also purchase a book. So going back to the earlier point I made that sometimes they can’t get through the book in a lending period, we’re seeing fully one-third of people are buyers as well as borrowers, so it’s a great audience to get in front of. There’s millions, hundreds of millions of page views at the library. Some people are looking through the title, the metadata and the information, doing searches. Sometimes they don’t have a library card to make the borrow at that time. Sometimes the book is checked out. So it’s a great place for discovery.

KENNEALLY: Right, exactly. And one way that libraries do that is to offer excerpts of books. And again, you’ve found that that really is a very effective tool for engagement and for commitment.

BURLEIGH: Right. So we provide that. They’re called OverDrive samples. And so you can click on the book and just read the first 10% of the book, and we do that in a couple of different ways, not just in the library but outside the library. Most publishers provide their approval to do that, but we see that, if you don’t do that, there’s a five – I’m sorry, people who do provide the excerpt to us, increases their title sales by up to five times. So we see that, in the li – and this is the B2B sales, so into the library that we can track the sales there.

KENNEALLY: Right. And Jared Friedman, you’ve been at the subscription business now at Scribd for, what, about a year and a half. And you’re beginning to see some real revenue return to publishers. And in particular, you’re distinguishing two types of revenue. I think it’s important to address that with regard to this notion of excerpts. Tell us about that. I’m thinking about browsing versus whole consumption.

FRIEDMAN: Yes. So at Scribd, we pay publishers when users read more than a certain amount of a book, but we also pay for what we call browsing revenue, which is our idea for how to give publishers a return on reading activity that previously has gone un-monetized. So if you’re just wandering around in a bookstore, I’m sure you’ve all had this experience, you’ll pull a book off the shelf, you’ll leaf through a few pages, you’ll return the book. Even though you’re looking for a book to buy, you’re still gaining some value from the sort of experience of browsing. And so we felt that there should be a business model around that sort of activity.

KENNEALLY: Right. Michael Tamblyn, you want to sell books. You want to help your retail partners sell whole books. But excerpts are certainly a discussion point in the publishing business of great interest and great debate. Can you contribute to that?

TAMBLYN: Yeah. We’ve found excerpts is a kind of an area of study that we’ve dug into a bit. We’ve looked at different lengths, different sizes, different delivery formats. For us, the greatest single swing in whether or not sort of a preview or an excerpt is effective is how easily and how quickly can you get it in front of someone? Do they have to go over to an app to read it? Can they get it just in front of the Website when they go? And if they do, conversion rates go crazy.

The length of it doesn’t seem to matter as much. So we’ve experimented with 10%, 15%, 20%, 30%. Somebody who, to Micah’s point, if you can get 10% of the way in, you’ve got some momentum. You’re likely, from our perspective, to become a purchaser. If we give you twice at much, it doesn’t make you much more or less likely to continue on as a purchaser.

KENNEALLY: Micah Bowers, you’re less concerned about purchases, but yet I think what Bluefire offers is important to this audience because it does talk about discovery. And that is where people are getting their material. They can find their material in a typical sales channel, but they can also get it from various public domain sites. They can import from a variety of services. I think you call it sideloading into the app. Describe how that works and what you find there. What are the sources that people are going to for their books?

BOWERS: Well, it’s really quite broad. And really, from the use-case standpoint with our applications, users can either just, what I call sideloading, you literally take a file that they have on their PC. They load it into the reader app on their mobile device, and they read it that way.

And people tend to load quite a few files. I mean it’s interesting, oftentimes we’ll see people load 100 files, 300 files, because – and then that’s interesting, because you know that people aren’t reading all those things on the device. And so it really gets back to this idea of people just like having their library with them. And so that’s kind of an interesting thing on the sideloading side of things.

On the Web links side of things, in our applications, people can be on another Website, click on a link. If it’s to an ePub file or to a DRMd pub file of certain types, it will just say, would you like to open this in an app? Which app would you like to open it in? Ours will show up. And they will open it. And that’s a link-based acquisition.

And we actually find, I think I actually have here some interesting stats on the acquisition patterns. There we go. And then there’s a third method in our particular case, where we have a bunch of library public domain books, available inside the app that people can download. So in our particular case, of the content that’s loaded on people’s devices, a third, 33%, will come from people sideloading the files, 65% come from people loading those from various Websites all over the Web directly into the app. And 2.5% are the free books that people can read in the app. And that was kind of surprisingly low to me. I mean there’s several thousand public domain books that are available inside the application. I would have thought that would be a much more common use case, but it’s 2.5% of the actual consumption.

And in our case, a lot of books are coming from that 65% of the Web link acquisitions. That means someone’s on a Website, there’s a link to a book, they can just click that and begin reading it inside the application. And now it’s there, downloaded into the app for reading later on, when they’re offline. But that’s coming from, in our case, it’s a mixture of public libraries, a lot of retailers, and then there’s just sites all over the place where, if you’re going to your employer and they have ePubs or PDFs that you need to download, so I mean it’s a broad perspective.

I don’t actually have a way of categorizing the types of sources because, from this perspective of the app, it’s just it was a link, and it caused the app to load. So it’s always one of those things that are a great mystery to me. When we look at hundreds of thousands of files a day coming into this app, we can see the title of the book, we can see how it’s read, what’s highlighted, what’s shared and all that, but we don’t know where it came from, which is kind of an interesting position to be in.

KENNEALLY: Indeed. We were talking about use cases, and it seems to me we’ve been discussing really the adult marketplace. But David Burleigh, libraries are a tremendously important source for parents of reading material for their children. And I’m thinking of illustrated books and other types of books that are of interest to the younger crowd. What do you know from the OverDrive app, sorry, the OverDrive application, that is telling us about the consumption of those types of books, illustrated books, enhanced books?

BURLEIGH: Yeah. This year, in 2014, OverDrive provided the ability to publishers to provide us with ePub 3 files, HTML5, so we can do enhanced e-books for fixed layout and, we call, audio sync text (sp?) or narrated e-books. So we have a few thousand of those now in our catalog. We’re adding more. These are like the Dr. Seusses of the world, and Chronicle Books has a number of these books as well. And so we’re seeing great pickup in the library. Obviously, this is really what people have been asking for for a long time, some synchronization of an audiobook and the visual e-book text that you can see in front of you, and you can follow along with it. It’s not an app. It doesn’t have to be built in multiple different operating system platforms. But it’s EPUB. It’s an EPUB 3 file that you can then play and put it into the channel for libraries and schools. We use our OverDrive Read browser-based reader, so it is a Webpage, and you can open it in any browser. If you can do that in that type of format, that will increase sales to the libraries – again, this is B2B sales, not consumer sales – but it increases sales versus the average book by about 12 times. So we’re seeing a lot of interest, a lot of excitement around it.

KENNEALLY: Michael Tamblyn, how much is the Kobo e-reader a family device rather than a personal device?

TAMBLYN: Most devices tend to be used by an individual person. But we have this interesting hand-me-down effect. New devices come out all the time, so what you’ll see in the data is someone gets a new device, that old device doesn’t tend to die. It goes to somebody else in the house, who then starts up their own stream of reading, which is an interesting challenge for us, because then they may be on essentially a family account, but we’re having to then sort of break into two streams of personalization and two streams of recommendation instead of just one. And that’s also quite different country to country. The farther down the road of digital adoption a country has gone, the more likely you’re going to see multi-device accounts in sort of the same house.

KENNEALLY: What about that, Jared Friedman, do you see that family device habit? And what does the data reveal to you on that?

FRIEDMAN: We do. So in your first question, you asked about one person using multiple devices, but there’s an interesting sort of inversion of that question, which is one device, multiple people using it. In our case, we’ve looked at it primarily multiple people using it simultaneously as opposed to the hand-me-down effect, which is also interesting, but more the effect of tablets that are being shared within a family.

So Netflix, for example, has these family accounts. And we can certainly see why they went down that road. Particularly with the rise of Android tablets in the past year and a half, we’ve seen a lot of adoption of these inexpensive Android tablets, we’ve seen that a lot of them are being used in a shared setting, where the adults read them at night and the kids read with them during the day.

KENNEALLY: Yeah. Michael Tamblyn, are there differences in that, you mentioned sort of the developing world and so forth, I wonder if we can talk about languages and how devices reveal language use habits and so forth. I believe, in your experience, you’ve seen the initial spike is for English but then, when the local language is introduced, English falls away.

TAMBLYN: I think some of the effect that Jared’s talking in terms of seeing strong English language sales, in especially some developing markets or markets that are just getting started in digital, is really about the kind of dearth of local language material.

So we walked that same pattern all the way through Europe and now into Asia, where you would enter into markets where 20%, 30%, 40% of access was in English, but that was because the publishers really hadn’t fired up their own conversion programs. They hadn’t made their own local titles available. Popular local authors weren’t in digital yet. And as soon as those gears begin to mesh, as soon as those publishers start making those titles available, English always kind of collapses down to usually kind of single-digit baseline thereafter.

And we pulled apart at that enough to see that really what we were picking up at the beginning of a digital adoption cycle were kind of transplants, expats, people who were trying to reach back to English language content from the country that they were in, which was not an English-speaking country, and then the next wave of people we were getting were local language readers who were actually really passionate about their own language, surprisingly.

KENNEALLY: Right. And David Burleigh, libraries again are a tremendous source, particularly for the immigrant population, to get materials from the home country, if you will. Does OverDrive offer its services in more than English?

BURLEIGH: Oh, absolutely. We supply libraries and schools in about 50 countries. The vast majority of our sales and our delivery of service is to the U.S. and to the English-speaking world, like the U.K. and Australia, Canada. But we’re seeing a lot of interest now in bringing Chinese, Japanese, Russian, Portuguese e-books, either translations of English material or books in their native language into North America. And we’re seeing a lot of sales into libraries to serve their largely urban populations, but really anywhere where there’s a pocket of those speakers in their communities, the libraries want to serve them.

KENNEALLY: Right. And because the libraries see what you offer as a kind of virtual branch, if you will, I’m sure that’s increasingly important to them, because it’s an offering that doesn’t require those people who are perhaps new to the community, may not be in the library-using tradition but definitely have personal computers and so forth, it may really welcome them in a way that an actual physical brick-and-mortar branch may not.

BURLEIGH: Right. Definitely. I mean the virtual branch is the perfect way they describe themselves. And in many cases, libraries will quantify their circulation across other branches, and they’ll consider their e-book library as a specific branch that out-circulates 80%, 90% of their physical branches. So they’re trying to, you know, this is one way libraries have been doing it for years, but they’re always looking for ways to reach more of their communities.

KENNEALLY: Right. Micah Bowers, let’s turn in the last few minutes, before we go to questions, to again this notion of moving beyond measurement and using what we are learning from the data to actually make differences in our business practices. Talk about that. How prepared do you feel publishers are to seize this opportunity? I know you’ve just begun to offer your analytics to some publishers and to others in the publishing business.

BOWERS: Yeah, I think there is definitely a lot of interest among publishers for this information. I mean it’s really quite compelling, actually. And a little personal anecdote is, you know, now that I actually have access through our system to this analytic data, it’s really quite addictive. I mean you can sit down, and you can be reading a book and say, gee, I wonder how other people are reading, and be able to go in and see that. So anyone who loves books, it’s incredibly compelling. And I often spend all night long just looking at the stuff.

But so there’s definitely – we’re early on, and there’s a long ways to go for publishers, and retailers, I would argue, but, to really make that kind of data actionable. And that really comes down to we’ve got a several hundred year old industry that is only beginning to be able to have access to this kind of data in the last, let’s say, year. And so it’s incredibly new. And it’s going to take changes within the organization, because this kind of data and the impact it can have across, whether it’s marketing and search engine optimization or on content structure or on marketing decisions or who you give that next contract to because of the engagement rate of – when you’re choosing who to give that advance to, looking at the analytic data, all of the things that publishers do can be highly informed by this analytic data.

And so there’s a long ways to go for organizations to make that happen. And in a lot of ways, it’s not just a matter of not having the processes and the job roles and the best practices and so forth, but sometimes it’s in fact counter to tradition and ego and all kinds of things that relate in the publishing industry around, you know, just there’s times when there was a magic to these decisions that now data might take a little bit of the magic out of. And that can create fear within organizations.

KENNEALLY: Right. Michael Tamblyn, you’ve been taking the magic out of the mystique for some time with your offering for BookNet Canada and that sales data offering. That goes back to the late 1990s. Are you finally seeing – and that was – I should say, it was revealing information, but it was revealing in kind of a two-dimensional way. The kind of data that we’re getting today is three dimensional, and that makes a difference. Is there the same passion for this data that you may have seen in the late ’90s, when you introduced the BookNet services?

TAMBLYN: Well, I think there are parts of it that are the same, and there are some bits that are quite different. I think the thing that is the same is that, like the ’90s, when point-of-sale data was starting to move into publishers, there’s apprehension institutionally because decisions can now be evaluated in a different way. And that’s always a bit anxiety causing for some people, who are worried that their decisions may not be looked on favorably.

But what we look at it more as is that, up until now, you’ve had two ways to decide whether a book was successful. You had the critical response, and you had the unit sales that sort of moved out the door. And this kind of brings the reader into the picture. Instead of just the salesperson and the marketing person and the performance of a retail channel, once it got in people’s hands, did they get excited by it? Were they engaged with it? Did they finish it? And I think one of the themes that comes through this whole last couple of days is we are more and more trying to bring the reader into the center of the discussions that we have around the books that we put in the market, and this is a great way to do it.

The thing that’s good about now versus, say, 10 and 15 years ago is publishers now have big data organizations, they have quantitative analysts. There are existing skill sets that are much more likely to be in place that can take advantage of it. We’re not trying to start this whole thing from scratch.

KENNEALLY: Right. Jared Friedman, I was thinking, as Michael was talking, about critical estimation of a book and the sales estimation of a book, you’ve got some interesting numbers on a particular author, Kurt Vonnegut. And I’m imagining that there could be a Ph.D. thesis in some of those numbers, or certainly it will contribute to an evaluation of Vonnegut in a way that wouldn’t have been possible before. You’ve looked at the reading behavior on various titles across Vonnegut’s career. Tell us about that.

FRIEDMAN: Exactly. So what this was motivated by was a goal that we had to go beyond just looking at a book, and to Michael’s point, saying is this book successful or not, but to actually taking it to the next level and looking inside the book and see if we can figure out why or why not, the sort of information that would be helpful if it could be handed back to the author, so that they could learn something from a particular book experience, and perhaps improve for their next book.

And so what we did is we took a bunch of popular authors, and we looked with who – sorry, a bunch of popular authors who had authored a considerable number of books, and then we looked across their books at the reading behavior. How quickly do readers read each of the books? Is it the same? Is it different? Has it changed as their writing style has developed? We look at the falloff rates. Are readers falling off? And if so, where? Is there a particular point where the author seems to lose the reader? And I think that a really compelling opportunity for the future is to get this data flowing back to the publishers and authors, who are going to be able to do something with it. For Scribd, it’s interesting but not really actionable. We aren’t going to rewrite the books.

So for Kurt Vonnegut in particular, he was one of our case studies, we found that there is in fact an enormous difference between the reading speeds and completion rates of his books. So Cat’s Cradle, for example, which I’m sure a lot of you have read, has a very fast reading speed, a very, very high completion rate. It’s known to be a very digestible book. But he has a very experimental book which he wrote, which jumps between different time periods and weaves a lot of story lines together and has a much higher number of characters, and we find that readers actually read that book about half the speed, and the completion rate is less than half. And that, I think, really does sort of prove out the hypothesis that it is possible to correlate behavior that we see on the reading side with the writing style of the author.

KENNEALLY: Right. Well, I was thinking of, you know, what would Vonnegut’s reaction be? He wouldn’t say so it goes. He would probably say to hell with that, I’m going to write the book I want to write. We have a few minutes for questions for our panel. If you have a question, let us know who it’s for. Yes?

M: Excuse me. (inaudible).

KENNEALLY: Oh, great. Great. If you wait for the microphone,

M: It really can be for anybody who knows the answer, because I’m a publisher, a travel publisher, and part of the (inaudible), we have books with all you guys. And we absolutely want to act on this data. It sounds to me like a job for Nielson (sp?) but, you know, because I don’t want to be going silo by silo and kind of (inaudible). So is Nielson thinking about this? What’s the plan?

M: I have no doubt, if there’s a data point out there that Jonathan Nolan (sp?) can wrap his arms around, he would love to do so. To be honest, I don’t think we’re there yet. I think we’re all still learning about the best way to measure – you know, all of us, I think, are struggling, for example, even with detecting the right place to figure out where books actually end and where they begin.

And we’re trying to figure out which metrics are most relevant to the publishers that we’re engaged with from a decision-making standpoint. It certainly isn’t kind of a raw stream of numbers that are just dropping back the way the point of sale does. So I think we’re still early days on this.

KENNEALLY: Micah Bowers, what do you think?

BOWERS: Well, I completely agree that it’s early days. And I do think that what you’re discussing, not just in the case of an aggregation of data, but it also speaks to the fact that, as we move forward, we’re going to need some kind of common language to talk about it. And so there’s multiple datasets. And how do you integrate that data from multiple sources? And it’s not just about who does that, but about what do we call things? And so I think there is work to be done.

And there was a meeting this summer, a BISG event that was an executive lunch that Kobo spoke at and that we spoke at as well. And one of the topics that we touched on there was the topic of a need for the beginnings of conversations across the industry on some normalization of, let’s say, nomenclature. And I’m not talking yet quite specifications, although we probably need to get there at some point. But I think even beginning to have the conversations of what do we call these things, and Michael brought up what do you consider to be the end of the book, is a reasonable thing to discuss.

KENNEALLY: Right. And it’s funny. If we were playing Digital Book World bingo, you just said something, I think, that probably comes up a lot, standards. And that’s going to be very critical moving forward. We have time for one last question, if there is, from the audience. Yes, back there, please? And wait for the microphone. And who is the question for?

F: It’s for the entire panel. I’m a lawyer for the Association of American Publishers, trying to represent the best interests of the industry. And so I’m here essentially to hear why data is so important to the publishers and all the different ways you want to use it.

It sounds like there are so many different ways that it could be used, and we’re still in the very early stages of it, but there are also conversations in Washington now about reader privacy and trying to make sure that only the limited amount of information and data is collected from readers that is absolutely essential to them. But I don’t think it sounds like we know what is essential to collect or not yet. So my question for all of you is, in an upcoming privacy discussion, what would you be afraid is going to be limited from data collection that you haven’t even yet explored the full potential of? And what does the industry association need to make sure we keep room for in order for the industry to thrive with this new possibility?

KENNEALLY: A really important question. And thank you for that. And David Burleigh, you’re dealing with the public library system, and people who go to those public libraries probably – they must know that the library, at least, knows what books they’re taking out. But they may not want that shared with anybody else. The kind of data collection that we’re talking about here, I think all of the panelists are talking about, is aggregate data. No one knows particularly what book Chris is reading, what book Michael is reading. But address that concern. Where does privacy play in all of this?

BURLEIGH: Privacy is huge. It’s paramount to the foundation of a library, so we certainly support that. Any data we collect is at the aggregate level. It’s not personally identifiable. But you’re right, I mean there’s going to be a range, or where do you draw that line? And so we’re trying to find out, just like everybody else, what information is important to us, what information is important to the publisher, to the library? What can we share? What is proprietary, based on the system? I mean there’s a lot of questions to sort through. But it’s absolutely the most important thing to preserve confidentiality.

KENNEALLY: Absolutely, and a great place to end this discussion today, and learning more about how they know what you read. I want to thank our panelists, Micah Bowers, with Bluefire, David Burleigh with OverDrive, Jared Friedman from Scribd and Michael Tamblyn from Kobo. My name is Chris Kenneally, with Copyright Clearance Center. Thank you all very much indeed.


Share This